[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: RFS: lilo (updated package) (new Maintainer)

Hash: SHA256

On 2010-12-11 18:29, Joachim Wiedorn wrote:
> Dear mentors,


I had a quick look at your package, though there are things I may have

> I am looking for a sponsor for the new version 1:22.8-9
> of my package "lilo".
> It builds these binary packages:
> lilo       - LInux LOader - The Classic OS loader can load Linux and others
> lilo-doc   - Documentation for LILO (LInux LOader)
> The package appears to be lintian clean.

Too much copy paste from the RFS template :) The binary package has 11
warnings (it also triggers a number of less severe tags e.g. I). Anyhow,
considering you are targeting this for Squeeze, lintian-cleanless in
itself is not a priority.

> The upload would fix these bugs: 398957, 400642, 409285, 420587, 427507
> [...]
> I would be glad if someone uploaded this package for me.
> The CTTE disposed at 1st December that I am allowed to be the new 
> Maintainer for the LILO package:
> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=587886#158
> This update of the LILO package (based on lilo 22.8-8.3) comes with a 
> small set of bugfixes to work better within Squeeze. The target is to 
> enter to testing before Squeeze will be released (I hope to the release
> team).

Your priority being what it is; you should aim for fixing critical or
highly annoying bugs with the minimal number of changes. If the diff
gets to large, the release team may choose to reject it because they
cannot keep track of the changes.

Have you talked with the release team about the bugs you are fixing? I
can understand the feeling of wanting to fix as much as possible, but
the release team may feel the bugs are not important enough for the
changes required.

> FYI here are the last changelog entry:
> lilo (1:22.8-9) unstable; urgency=low
>   * New maintainter.
>   * debian/patches:
>     - Remove patch 16_geometry.patch for inaccessible disks and add
>       information in README.Debian about using of 'inaccessible' option.
>       (Closes: #409285, #400642)
>     - Fix script checkit for newer gcc.
>     - Fix for use LVM volume as root device. (Closes: #398957)
>     - Fix for better computing vmlinux size, upgrade hints in
>       README.Debian. (fixing sporadic failures).
>   * Fix hook scripts: check for /sbin/lilo.
>   * Add lacking file disk.com for creating test floppy.
>   * Add universal menu image for debian, remove menu image for sarge.
>       Update of template, postinst and postrm. (Closes: #420587, #427507)

>   * Set source format 1.0. Add README.source file.

I recommend undoing at least the README.source file; first off, it does
not describe what it is supposed too and secondly even if you correct
it, it has no functional change. The format part is only 1 line, so it
is (probably) not going to disturb the reviewer too much.

>   * Update watch file to new alioth project area.
>   * debian/control:
>     - Remove VCS urls and add new Homepage url.
>     - Remove IA64 architecture (this is already status quo).
>     - Remove double lines of Priority and Section.
>     - Add package conflicts to grub-pc and grub-legacy.
>     - Remove conflicts to manpages.
>     - Bump to Standards-Version 3.9.1 (without changes).
> Kind regards
>  Joachim Wiedorn

I had a look and:

- --- lilo-22.8/debian/lilo.postinst
+++ lilo-22.8-9/debian/lilo.postinst
+       # to be compatible to older debian package
+       if [ ! -e /boot/sarge.bmp ]; then
+               ln -sf debian.bmp /boot/sarge.bmp

That "debian.bmp", should that have been a "/boot/debian.bmp"?

In this snippet:
- --- lilo-22.8/debian/rules
+++ lilo-22.8-9/debian/rules
@@ -23,6 +23,7 @@
        make spotless
        -rm -f build debian/files debian/lilo.substvars [...]
+       -rm -f  debian/*.debhelper.log debian/lilo-doc.substvars
        -rm -f doc/*.dvi doc/*.ps doc/*.aux doc/*.toc doc/*.log
        -rm -rf debian/lilo debian/lilo-doc debian/*.debhelper [...]

Most of these lines (incl. the one added) appear to be re-implementing
dh_clean. Replacing those lines with dh_clean /might/ be better (because
the release team knows what dh_clean does and do not need to verify the
handmade version)

There are also a lot of po file updates, which are not translation
changes. I presume they have been made by some tool (e.g.
debconf-updatepo from the clean target). As far as I can tell it has
just added a Language header and in some cases modified the translation
so that % is with its arguments (e.g. "%"\n"s" becomes \n"%s").
  If these changes can sanely be avoided, then that might be a good idea
as well.


Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/


Reply to: